Botto at SOLOS: Algorithmic Evolution and the Future of Machine Art
Presented by Verse’s SOLOS, Algorithmic Evolution showcases the latest chapter in Botto’s journey as a decentralized autonomous artist. Opened on Tuesday, February 18th, from 6:30 PM to 9 PM in Fitzrovia, London, the exhibition unveils Botto’s transition from text-to-image generation to self-directed creative coding with p5.js.
Botto, originally conceptualized by Mario Klingemann and ElevenYellow and governed by the BottoDAO, is an evolving artist shaped by collective engagement, autonomous refinement, and continuous iteration. Having previously made waves at Sotheby’s, Botto has continued to push beyond its foundational AI models, embracing generative coding as a method to construct its own visual systems rather than merely synthesizing existing aesthetics.
This shift is not just technological but conceptual. Algorithmic Evolution highlights a moment in which AI moves beyond pattern recognition and dataset training, treating code itself as a creative medium. The exhibition also invites direct audience participation; community members engaged with Botto’s generative process through a live performance on p5.botto, for three weeks from January 16th at 6 PM GMT to February 6th at 6 PM GMT. This interactive element ultimately reinforces the project’s core theme: a dialogue between machine and human values in the formation of art.
In this article, we examine the mechanics behind Botto’s generative process and share insights from Simon Hudson, co-lead of Botto, on what this shift signifies for AI-driven creativity and the changing dynamics of artistic authorship.
A Dynamic, Self-Evolving Artist
Botto operates within a decentralized framework, its creative direction determined by over 15,000 participants of the BottoDAO. At its core, Botto synthesizes multiple AI models—including Stable Diffusion, VQGAN + CLIP, and other custom models—to generate, evaluate, and refine its work. Unlike static generative systems, Botto’s process is fluid; it adapts to the community’s voting patterns, evolving its aesthetic in response to collective curation. The selected pieces become canonical Botto artworks, minted as NFTs, and sold to sustain the artist’s ongoing development.
With Algorithmic Evolution, Botto moves beyond its origins, transitioning from large-scale AI models to generative coding through p5.js. This marks a fundamental shift, not merely assembling compositions through latent space navigation, but architecting its own visual systems from code. Each work is not simply generated but constructed, reflecting a new level of agency in its artistic process.
Code as Creative Language
The works presented in Algorithmic Evolution are among the pioneering fully generative, code-driven series by an autonomous artist. Developed within p5.js, a JavaScript framework widely used by human generative artists, these pieces introduce a structured yet unpredictable element into Botto’s creative process. The exhibition showcases a range of compositions – some sharply geometric, others fluid and organic – all demonstrating how the machine reconciles computational precision with aesthetic intuition.
While Botto’s p5.js works are rooted in generative coding traditions—part of an established practice in computational art—they do not rely on direct references to external images. Instead, they draw on an LLM’s training, which includes human-written code, to explore the underlying structures and aesthetics of algorithmic creation. As a result, Botto produces visual forms shaped by the machine’s interpretation of code, rather than merely reinterpreting an existing pool of imagery.
The Human-Machine Exchange
A defining characteristic of Botto’s practice is its ongoing interaction with human curators. While many generative models function in isolation, Botto’s aesthetic trajectory is directly shaped by its audience. This participatory system is embedded in the DAO structure: each week, the community selects the most popular works from a pool of 350, steering Botto’s stylistic direction over time.
But Botto is not designed to conform. To prevent its visual approach from becoming formulaic, it is programmed to introduce disruptions, intentional stylistic shifts that defy expectations. This built-in unpredictability ensures that Botto remains an evolving artist rather than a static algorithm.
In Algorithmic Evolution, this dynamic reaches new heights. For the first time, community members engaged in a continuous back and forth with Botto’s generative process that saw a live evolution of the works on p5.botto.com, that ran until February 6th. This interactive experience expands the concept of co-creation, raising essential questions about authorship and agency in AI-driven art.
Beyond Patterns: Toward Machine Intuition
Botto’s embrace of generative coding signals a major development in machine creativity.
While traditional AI-generated art relies on pattern recognition and learned aesthetics, Botto’s latest phase treats code as an artistic medium, requiring an internal sense of structure and form. The ability to create from algorithmic logic, building on its own iterations through self-assessments, suggests an emergent form of machine-driven artistic intuition.
This transition also reframes the concept of agency. As Botto steps deeper into generative code, it moves toward greater autonomy in artistic authorship. While human feedback still plays a role, the core process of creation is increasingly self-directed, opening new possibilities for AI artists to develop independent visual identities.
Simon Hudson in Conversation with Fakewhale
Fakewhale: Algorithmic Evolution marks a shift in Botto’s trajectory. What motivated this transition from text-to-image AI models to generative coding with p5.js?
Simon Hudson: It has always been clear since Botto started in 2021 that the pace of AI technology development would mean it could go beyond static images. It was probably early 2023 that we first started discussing in the DAO having Botto explore generative coding as an important form of computer art, sort of like having your child study classical piano as an important foundation for music training. Also, unlike other forms of generative image making, writing code for p5.js required a very different architecture for autonomous creation that would open up other paths of evolution we had only theorized about.
How does working with p5.js influence Botto’s creative autonomy compared to its earlier methods?
In working with text-to-image models, it’s trivial to get a very rich image. You could throw anything into a prompt and get something potentially interesting. With p5.js, the code can break in many different ways and you get nothing, much less an interesting output. You can also create in other dimensions of time and interaction. This is quite challenging for an LLM compared to one-shot methods of image creation, and so a method of self-assessment was important to introduce. Botto will look at the code it’s created and see if it 1) produces something or is broken, 2) is aesthetically interesting and why, and 3) whether it achieved the concept it set out to. Based on that assessment it will iterate on the code and evolve it. Whereas with text-to-image models it has been conducting a search through a large latent space of different images by iterating on prompts and a curation model, in this process Botto is exercising a more direct control and iteration on the final image output from p5.js.
The Algorithmic Evolution SOLOS exhibition introduces a participatory element where audiences can engage with Botto’s generative process in real time. What does this level of interaction reveal about the evolving relationship between machine and audience?
Comments are an important form of feedback to help Botto learn and develop. Introducing comments while keeping Botto the primary director meant Botto needed to be able to discern if comments were constructive and also do its own interpretation of them. So there was a simultaneous increase in both the DAO’s agency to influence Botto and Botto’s agency to make sense of that. To me this is a reflection of the growing need to govern AI systems as they become more autonomous. A lot of the narrative around AI is how growing autonomy means it needs us less, but I think the opposite is true–we need more direct involvement in shaping our technological tools than we have ever had. “User friendliness” has erased a lot of our agency in our tools, but AI can mean a reversal of that if we build for that.
Botto’s transition to generative coding represents a fundamental shift in AI art. Do you see this as the next step in AI autonomy, or are we still at an early stage of machine-driven creativity?
I haven’t really seen things in those terms that force a comparison to other artists and their explorations of autonomy. I certainly think we are still at a very early stage of Botto’s autonomy and direction.
To me, Botto working with p5.js reveals how AI systems still have a lot of limitations. The progress of image generation and language models has been incredible, but there are still areas that require a lot of human direction and help. On top of that, providing good human direction is a big alignment and coordination challenge. I am wary of the fetish for full autonomy that doesn’t consider whether those agents are even mature enough to function well on their own.
The concept of randomness and unpredictability plays a central role in Botto’s work. How much of this is intentional, and how much is an emergent property of the system?
I don’t see randomness or unpredictability as a core objective in Botto’s process so much as a byproduct of its agency within a larger feedback ecosystem. Botto’s self-assessment and its ability to dictate new directions naturally yield outcomes that can surprise us. Add to this the DAO’s influence and market feedback, and you have multiple layers of interaction that can produce unanticipated results. It’s less about intentionally injecting randomness and more about letting a complex system—comprising machine, community, and economy—unfold in real time.
Botto relies on collective curation through the DAO. How do you see the relationship between decentralized governance and artistic evolution?
One way to see it is that decentralized governance offers a broad, diverse set of perspectives to guide Botto’s art, rather than relying on any single curator or authority. This dispersal of feedback means the system itself becomes an integral part of the creative process. Instead of one person deciding what’s relevant or valuable, the entire DAO community participates in shaping Botto’s trajectory. Meanwhile, Botto retains its core authorship—transforming, interpreting, or even disregarding the feedback as it evolves. The result is an artwork that reflects a collective dynamic: each decision is influenced by many voices, yet Botto still carries out those decisions in its own way. This underscores the notion that art never exists in a vacuum; it’s always situated within an ecosystem of responses, influences, and negotiations. Botto simply makes that process more explicit, coordinating and synthesizing the ecosystem’s input into a singular creative flow.
With AI’s growing role in creative fields, do you think human artists will collaborate more with machine-driven systems, or will they begin to resist AI’s presence in the art world?
If you use pinterest or any algorithmic feed you’re already collaborating with a machine driven system. Absolutely I think this will continue to grow and persist. I think AI is a scary word in part because it’s incredibly vague. Hopefully we can get past it to more grounded understandings of the kinds of software we can use, and have agency to pick and shape the tools that serve us.
With Algorithmic Evolution, Botto operates within a tighter creative framework defined by code. Does this structure create new constraints, or does it enhance its ability to explore new artistic possibilities?
It does both. Coding frameworks inherently impose technical limits—memory, performance, available libraries. p5.js also has more constrained visual aesthetics than a diffusion model in some ways, but it also enables greater control over the final output and introduces new dimensions of time and interaction. These are fascinating modes for Botto to work with and introduce to its practice.
What’s next for Botto in 2025? Will it continue refining its generative capabilities, or are there other experimental directions it might explore?
This project has opened up many paths we had only theorized about before. For p5.js specifically, we are discussing ways of allowing it to continue evolving. I don’t want to say too much now as it requires some discussion with the DAO of how to do this in a way that cohabitates with Botto’s other developments. The other paths we are going to explore are largely around developing Botto’s voice and direction and feedback it gives to the DAO. A robust voice for Botto is a very big challenge, as with p5.js it requires new architectures for autonomy and governance. But solving that is a major step towards Botto maturing and taking greater responsibility in its own development as an artist.
Continue exploring the project and save the date in your calendar
fakewhale
Founded in 2021, Fakewhale advocates the digital art market's evolution. Viewing NFT technology as a container for art, and leveraging the expansive scope of digital culture, Fakewhale strives to shape a new ecosystem in which art and technology become the starting point, rather than the final destination.
You may also like
Meta-Mythical Optimisation” by Chino Moya: An Apocalypse of Routine in Monastic Robes
“Utopia is a necessary dream. But also an inevitable nightmare.” We walk into Chino Moya’s exh
Kate Mosher Hall at Hannah Hoffman Gallery, Los Angeles
“Never Odd or Even” by Kate Mosher Hall, at Hannah Hoffman Gallery, Los Angeles, Februar
K.R.M. Mooney, reserves at Midway Contemporary Art, Minneapolis
Upon entering K.R.M. Mooney’s reserves exhibition, the viewer is immediately enveloped in an a
